Fax: (713) 799-8712
Caveolin-1 expression is a predictor of recurrence-free survival in pT2N0 prostate carcinoma diagnosed in Japanese patients
Article first published online: 20 FEB 2003
Copyright © 2003 American Cancer Society
Volume 97, Issue 5, pages 1225–1233, 1 March 2003
How to Cite
Satoh, T., Yang, G., Egawa, S., Addai, J., Frolov, A., Kuwao, S., Timme, T. L., Baba, S. and Thompson, T. C. (2003), Caveolin-1 expression is a predictor of recurrence-free survival in pT2N0 prostate carcinoma diagnosed in Japanese patients. Cancer, 97: 1225–1233. doi: 10.1002/cncr.11198
- Issue published online: 20 FEB 2003
- Article first published online: 20 FEB 2003
- Manuscript Accepted: 28 OCT 2002
- Manuscript Revised: 27 SEP 2002
- Manuscript Received: 18 JUL 2002
- National Cancer Institute. Grant Numbers: RO1-CA68814, P50-CA58204
- caveolin-1 (cav-1) expression;
- prostate carcinoma;
- prognostic marker;
The authors previously identified elevated caveolin-1 expression in human prostate carcinoma and determined that caveolin-1 levels as detected by immunohistochemistry of radical prostatectomy specimens offered novel prognostic information. A higher incidence of caveolin-1 expression also was reported in African-American men compared with white men in the U.S. To explore these ethnic/racial differences in caveolin-1 expression further, the authors evaluated caveolin-1 expression as a predictive marker in Japanese men with prostate carcinoma.
Immunohistochemical staining with a caveolin-1 specific antibody was performed on routinely processed paraffin sections from 152 consecutively collected radical prostatectomy specimens. The mean patient age was 64.3 years (range, 49–74 years; median, 64.5 years) and the mean follow-up period was 49.5 months (range, 1.3–103.3 months; median, 48.2 months). Caveolin-1 immunoreactivity was evaluated in association with patient's age; preoperative prostate specific antigen level; clinical stage; and pathologic features including Gleason score, extraprostatic extension, status of surgical margins, seminal vesicle involvement, lymph node involvement, and time to disease progression after surgery.
Positive caveolin-1 immunostaining was detected in 46 of the 152 tumors (30.3%) and was found to be associated significantly with a positive surgical margin (P = 0.022). A higher incidence of caveolin-1 expression tended to be found in patients with poorly differentiated tumors (Gleason score > 7, 6–7, and < 6, 35.0% vs. 34.9% vs. 20.4%, respectively) or in patients with extraprostatic extension versus those without extraprostatic extension (35.4% vs. 24.7%) or patients with lymph node involvement compared with those without lymph node involvement (50% vs. 29.5%), although these differences did not reach statistical significance (P = 0.100, P = 0.150, and P = 0.178, respectively, by the Spearman correlation test). Kaplan–Meier analysis revealed that increased caveolin-1 expression was associated with an increased risk of disease progression at 5 years (P = 0.0122 by the log-rank test). In patients with organ-confined (pT2N0) disease, univariate Cox proportional hazards regression analysis revealed that positive caveolin-1 expression was the only significant predictor of disease recurrence after radical prostatectomy (P = 0.011; hazards ratio = 4.75; and 95% confidence interval, 1.43–15.76).
The results of the current study confirm that positive caveolin-1 expression is associated with clinical markers of disease progression and is predictive of poor clinical outcome after surgery in Japanese patients with pT2N0 prostate carcinoma. Cancer 2003;97:1225–33. © 2003 American Cancer Society.
Prostate carcinoma is the most prevalent male malignancy in developed countries worldwide.1, 2 The incidence of prostate carcinoma has reportedly increased dramatically throughout the world in the past decade with mortality rates also increasing in many countries, although with considerable geographic variation. In Japan, there have been increases reported in the incidence of prostate carcinoma3 and the death rate with prostate carcinoma is believed to be the eighth most common cause of male cancer death in Japan.4 The incidence of prostate carcinoma was reported to be 10,940 in 1994 and this number may increase to as many as 30,285 by 2015.5
The increased incidence of prostate carcinoma is related to the widespread use of serum prostate specific antigen (PSA)-based screening of asymptomatic men. Although the incidence and mortality rates recently have begun to decrease in the U.S.,6 it is unclear whether these reductions are related to increased screening and earlier aggressive treatment, misclassification of cause of death, or more complex population dynamics.7–9 In addition, there most likely has not been adequate follow-up time in which to make meaningful comparisons of screening or treatment outcomes.10
The use of different cutoff limits for serum PSA in the detection of prostate carcinoma confined within the gland in the Japanese population has been proposed.11 Nomograms have been developed that utilize preoperative serum PSA, biopsy Gleason score, and clinical stage to predict the pathologic stage and probability of biochemical recurrence after radical prostatectomy in white men.12, 13 However, for cancer occurring in Japanese men or in the middle range of a given index, the predictive power of these indices is diminished greatly. The use of molecular markers to supplement clinical information concerning the biologic aggressive of a carcinoma may allow better selection among different treatment regimens.
Among the difficult problems associated with prostate carcinoma is the relatively high prevalence of histologic tumors with low malignant potential, estimated to be as frequent as approximately 40% in men age > 50 years,14, 15 which suggests that many of the tumors detected may in fact be “clinically unimportant.”15 In fact, data from some studies suggest that approximately 10–26% of nonpalpable tumors detected by serum PSA screening are “clinically unimportant” on the basis of pathologic criteria (e.g., < 0.5 cc, a Gleason score ≤ 6, and disease confined to the prostate).16–19 The detection and treatment of potentially “clinically unimportant” tumors with potentially harmful therapy such as radical prostatectomy and irradiation therapy may not be appropriate in many cases. Overall, the complex morphologic patterns, histologic heterogeneity, and the early manifestations of high malignant potential preclude a straightforward assessment of the metastatic potential of localized prostate carcinoma and indicate the need for additional clinical and pathologic tests for the objective assessment of prostate carcinoma stage and clinical/biologic potential.
We previously identified and characterized the overexpression of the caveolin-1 (cav-1) gene in a mouse model of prostate carcinoma and in human prostate carcinoma.20 Cav-1 is a major structural component of caveolae, and the plasma membrane and trans-Golgi network have been implicated in sphingolipid-cholesterol transport and specific signal transduction pathways.21, 22 Using a series of metastatic androgen-resistant mouse prostate carcinoma cell clones that were stably transfected with antisense cav-1, we demonstrated that reduced levels of cav-1 protein triggered apoptosis after the withdrawal of androgen both in vivo and in vitro.23 Reintroduction of cav-1 with an adenoviral vector in a representative antisense clone substituted for the survival effects of testosterone in vivo.23 In both mouse and human prostate carcinoma cell lines, we demonstrated that cav-1 is a downstream effector of testosterone-mediated survival activities and that modest but not high levels of cav-1 can promote both cell survival and metastatic activities in mouse and human prostate carcinoma cells.24 We also reported that cav-1 is regulated by c-myc and suppresses c-myc-induced apoptosis.25 Recently, we established that cav-1 expression is significantly increased in primary and metastatic human prostate tumors after androgen ablation therapy, that cav-1 is secreted by androgen-insensitive prostate carcinoma cells, and that this secretion is regulated by steroid hormones. It is interesting to note that cav-1 was detected in the HDL3 fraction of serum specimens from patients with advanced prostate carcinoma and, to a lesser extent, in healthy subjects.26 Therefore, cav-1 is a metastasis-related gene and a candidate gene for hormone-resistant prostate carcinoma in humans.27, 28
In cases of clinically determined, localized prostate carcinoma, we found that cav-1 expression is a novel prognostic marker with independent predictive value of biochemical recurrence.29 We also reported a higher incidence of cav-1 expression in African-American men compared with white men.30 To investigate ethnic/racial differences in cav-1 expression further, we evaluated cav-1 expression as a positive predictive marker in Japanese men with prostate carcinoma, including patients with pathologically determined localized disease.
MATERIALS AND METHODS
Patients and Prostate Specimens
One hundred fifty-two previously untreated Japanese patients with clinically resectable prostate carcinoma (T1-T3N0M0) who had undergone bilateral pelvic lymphadenectomy and radical prostatectomy at Kitasato University Hospital between May 1992 and June 2000 participated in the current study. Clinical stage was determined in accordance with the unified TNM system.31 All patients were seen once every 2 weeks for the first 3 months after surgery, quarterly for the first year, semiannually for 2 years, and yearly thereafter. Each follow-up visit required history, physical examination, and serum PSA. Rectal examination was conducted by one of the authors (S.E). Bone scans were obtained yearly. Chest radiographs were obtained or abdominal computed tomography was performed when clinically indicated.
Serum PSA was quantitated by the Eiken polyclonal radioimmunoassy (Eiken, Tokyo, Japan) until March 1993, with the Dainapack IMx PSA assay (Dinabot, Tokyo, Japan) until June 1997, and with the AxSYM PSA assay (Dinabot) thereafter.32–34 In one patient, preoperative serum PSA data were not available. The follow-up of 11 patients took place at other hospitals. For uniformity, Eiken PSA data were interconverted as recommended by Machida et al.32 as follows: IMx PSA = 1.39 × Eiken PSA – 1.02. Results of the AxSYM PSA assay were not interconverted because they are considered to be virtually the same as those of the IMx PSA.34
Postoperative serum PSA values were considered elevated if values of ≥ 0.1 ng/mL were obtained on 2 consecutive visits 1 month apart. Disease progression was defined as biopsy-confirmed local recurrent tumor; evidence of malignant disease on a bone scan, chest radiograph, or skeletal radiograph; or elevated postoperative serum PSA. In all but two patients, radical prostatectomy specimens were examined by whole-organ step-section techniques, as described previously.35 In the remaining two cases, random multiple sectioning was performed and each section was stained with hematoxylin and eosin. Tumor grade and depth of capsular penetration were determined on the basis of criteria specified previously.35 Postoperative pathologic findings including status of extracapsular extension, seminal vesicle invasion, lymph node metastasis, surgical margin, and Gleason score were recorded by a single pathologist (S.K.).
Immunohistochemical staining was performed essentially as described previously.20, 29 Briefly, sections were deparaffinized, rehydrated, and then heated for 9 minutes in citrate buffer (0.01 M [pH 6.5]) in an 800-watt microwave oven for antigen retrieval. The sections were treated with 2% hydrogen peroxide for 10 minutes to inactivate endogenous peroxidase and blocked with 3% normal goat serum in 0.2 M phosphate-buffered saline (PBS) (pH 7.4), followed by incubation for 2 hours with rabbit polyclonal antibody to cav-1 (Santa Cruz Biotechnology, Santa Cruz, CA) at a dilution of 1:400 in PBS with 0.5% normal goat serum. Standard avidin-biotin complex (ABC) immunostaining procedures (Vector Laboratories, Burlingame, CA) were followed by incubation in 3.3′-diaminobenzidine/hydrogen peroxide solution. Antibody specificity was documented previously.20, 29 The immunostained sections were evaluated at a power of ×200 under a Zeiss microscope (Zeiss Inc., Thornwood, NY). For each specimen, only the index tumor was evaluated. The entire tumor area was scanned and an average of 20 ± 9.3 (mean ± the standard deviation) microscopic fields were evaluated. Positive cav-1 staining for a particular specimen was assigned if > 50% of the tumor cells in any microscopic field (real area: 0.0625 mm2) demonstrated cav-1-positive granular immunoreaction products in their cytoplasm. All specimens were evaluated without knowledge of the patients' clinical information.
The correlation of cav-1 immunoreactivity with the patients' clinical and pathologic variables was analyzed using the Spearman or Pearson correlation test. The predictive value of cav-1 and other clinical and pathologic variables such as preoperative serum PSA, extraprostatic extension, positive surgical margins, and Gleason score were univariately tested using Kaplan–Meier actuarial analysis36 and the log-rank test. In addition, multivariate analysis of cav-1 immunostaining and the postoperative pathologic markers was performed using the Cox proportional hazards regression model.37 The risk ratio and its 95% confidence interval (95% CI) were recorded for each marker. A P value < 0.05 was considered statistically significant in all the analyses. The Spearman correlation test was used to test the difference in the incidence rates of cav-1 expression in patients from different pathologic subgroups. All analyses were performed with SPSS statistical software (SPSS Inc., Chicago, IL).
Cav-1 immunoreactivity was detected in the cytoplasm of tumor cells in a granular pattern (Fig. 1). Although the proportion of cav-1-positive tumor cells varied within individual specimens, the validity of the staining reactions was confirmed by strong staining reactions characteristic of smooth muscle and endothelial cells within the stroma. Cav-1 reactivity was not detected in histologically normal epithelial cells in areas adjacent to the tumor.
Cav-1 expression in relation to clinical and pathologic features is summarized in Table 1. Among the 152 prostate carcinoma specimens, 46 (30.3%) were positive for cav-1 immunoreactivity. The percentage of clinical stage T1a-b, T1c, T2a-c, and T3a-c tumors with positive cav-1 immunoreactivity was 66.7%, 29.4%, 27.5%, and 33.3%, respectively. No significant association between cav-1 expression and clinical stage was found (P = 0.493, Spearman correlation test).
|Preoperative PSA (ng/mL)||0.203|
|Seminal vesicle invasion||0.351|
|Lymph node involvement||0.178|
When the preoperative serum PSA levels of the patients were stratified into 3 groups (< 4 ng/mL, 4–10 ng/mL, and > 10 ng/mL), the patient group with a PSA level > 10 ng/mL was found to have the highest rate of cav-1-positive tumors (24.0%, 27.4%, and 35.9%, respectively). However, no significant correlation between cav-1 expression and preoperative serum PSA level was found (P = 0.203, Spearman correlation test).
The cav-1 immunopositivity was found to be significantly higher in the patients who had a positive surgical margin (42.9%; n = 49) compared with those with negative surgical margins (24.5%; n = 102). A significant correlation between cav-1 expression and a positive surgical margin was found (P = 0.022, Spearman correlation test).
With regard to the other pathologic parameters, there was in general a trend toward a higher frequency of cav-1 positivity with the more advanced parameter but this finding did not achieve statistical significance (Table 1). Gleason scores of prostatectomy specimens were stratified into 3 groups: well differentiated (Gleason score < 6; n = 49), moderately differentiated (Gleason score 6–7; n = 83), or poorly differentiated (Gleason score > 7; n = 20). Although there was a tendency toward a higher incidence of cav-1 expression with a higher Gleason score, no significant correlation between cav-1 expression and Gleason score was found (P = 0.100; Spearman correlation test). The frequency of cav-1 expression was higher in those tumors with extraprostatic extension (35.4%; n = 28) than in those without (24.7%; n = 18); however, this finding did not achieve statistical significance (P = 0.150, Spearman correlation test). In specimens in which there was seminal vesicle involvement, the frequency of cav-1-positive tumors was greater (17 of 48 specimens [35.4%]) than in specimens without seminal vesicle involvement (29 of 104 specimens [27.9%]). However, no significant correlation was found between cav-1 expression and seminal vesicle involvement (P = 0.351, Spearman correlation test). Although 5 of 10 of specimens from patients with positive lymph nodes stained positive for cav-1 (50.0%), this limited sample size precluded significance (P = 0.178, Spearman correlation test) when compared with the patients with no evidence of lymph node metastasis (41 of 139 specimens [29.5%]).
Cav-1 as a Predictor for Time to Disease Progression
The predictive value of positive versus negative cav-1 staining status was evaluated using Kaplan–Meier actuarial analysis (Fig. 2). The follow-up time ranged from 1.3–103.3 months (mean, 49.5 months; median, 48.2 months). The mean recurrence-free survival time for the cav-1-positive patients was 42.7 months (95% CI, 31.3–54.2 months), whereas that for the patients with negative cav-1 expression was 68.6 months (95% CI, 59.4–77.7 months).
Univariate Cox proportional hazards regression analysis was used to evaluate positive cav-1 expression (P = 0.0140); preoperative serum PSA level (natural log transformed) (P < 0.0001); Gleason score (1) (6–7 vs. < 6; P = 0.0046), Gleason score (2) (> 7 vs. < 6; P < 0.0001), and Gleason score (3) (> 7 vs. 6–7; P = 0.0138); extraprostatic extension (P < 0.0001); seminal vesicle involvement (P < 0.0001); and positive surgical margin (P < 0.0001) as significant prognostic predictors of a shorter time to disease progression after surgery (Table 2). Age provided no prognostic value in this set of patients.
|Parameters||All patients (n = 141)||Stage pT2N0 patients (n = 61)|
|HR (95% CI)||P value||HR (95% CI)||P value|
|Caveolin-1a||1.96 (1.15–3.36)||0.0140||4.75 (1.43–15.76)||0.0110|
|Preoperative PSAb||2.38 (1.71–3.32)||<0.0001||1.32 (0.57–3.04)||0.5154|
|Gleason score (1)c||3.06 (1.41–6.63)||0.0046||1.27 (0.39–4.16)||0.6947|
|Gleason score (2)c||6.75 (2.82–16.15)||<0.0001||—||—|
|Gleason score (3)c||2.21 (1.18–4.15)||0.0138||—||—|
|Positive surgical margin||3.25 (1.90–5.55)||<0.0001||2.62 (0.68–10.16)||0.1629|
|Extracapsular extension||4.35 (2.33–8.13)||<0.0001||—||—|
|Seminal vesicle involvement||4.46 (2.61–7.63)||<0.0001||—||—|
|Lymph node involvement||3.67 (1.71–7.88)||0.0008||—||—|
Multivariate Cox proportional hazards regression analysis with these parameters from the same set of patients also was performed. Cav-1 adjusted for other parameters and pathologic parameters such as preoperative serum PSA level (natural log transformed), Gleason score, extraprostatic extension, seminal vesicle involvement, surgical margins, and lymph node status demonstrate that it is not an independent significant predictor of time to disease recurrence (P = 0.7794). Of all the clinicopathologic parameters, only seminal vesicle involvement (P < 0.0001), positive surgical margins (P = 0.0044), and preoperative PSA level (natural log transformed) (P = 0.0161) were found to significantly predict time to disease progression multivariately.
Kaplan-Meier analysis was also performed in a subset of patients with pathologically confined (pT2NO) tumors (Fig. 3). In this group of patients the mean recurrence-free survival time was 85.7 months (95% CI, 77.8–93.7 months) for patients with cav-1 negative tumors whereas in patients with tumors positive for cav-1 expression the mean recurrence-free survival time was 46.5 months (95% CI, 30.6–62.6 months). When the univariate Cox proportional hazards regression analysis was performed on those patients with pathologically organ-confined prostate carcinoma, positive cav-1 expression was found to be the only prognostic factor for time to disease progression with a hazard ratio of 4.75 (P = 0.0110) (Table 2). Preoperative serum PSA level (natural log transformed), Gleason score (6–7 vs. < 6), and positive surgical margin did not appear to offer independent prognostic value (P = 0.5154, P =0.6947, and P =0.1629, respectively).
Serum PSA screening has led to a dramatic increase in the diagnosis of prostate carcinoma as well as in the number of men undergoing radical prostatectomy in the past decade.11, 38 In the U.S. the number of newly diagnosed prostate carcinoma cases and cancer specific mortality appear to be declining; however, it is unclear whether these reductions are related to increased screening and earlier aggressive treatment, misclassification of cause of death, or more complex population dynamics.7–9 Furthermore, there most likely has not been adequate follow-up time in which to make meaningful comparisons of screening or treatment outcomes.10
The treatments currently used for prostate carcinoma are exclusively local treatments designed to ablate the tumor either surgically or with irradiation. The optimal use of these therapies requires accurate staging of the disease. Unfortunately, neither low to intermediate serum PSA levels (approximately 2–15 ng/mL) nor currently available clinical modalities are capable of accurate staging. In as many as 50% of surgically treated, presumably localized, prostate carcinoma patients the tumor already has extended beyond the confines of the gland when examined pathologically.12, 17, 39
Preoperative serum PSA levels, Gleason score, and clinical and pathologic stage have been shown to be useful in predicting clinical outcome.12, 13 However, in the middle ranges of a given index, these indices lose predictive power. Therefore, additional individual predicting methods are needed, such as molecular markers.
We previously reported that positive cav-1 expression was a novel prognostic marker for human prostate carcinoma29 and reported racial/ethnic differences in cav-1 expression between African-American and white patients.30 However, to our knowledge cav-1 has not been studied in Japanese prostate carcinoma patients. The purpose of the current study was to assess the possibility using of cav-1 expression as a predictive marker for Japanese prostate carcinoma cases.
In a large set (n = 152) of Japanese prostate carcinoma specimens, 30.3% (46 of 152 specimens) were determined to be positive for tumor cell-specific cav-1 immunoreactivity.
There was a positive correlation between cav-1 expression and surgical margin positivity and a trend with Gleason score also was found. In addition, a higher incidence of cav-1 expression tended to be found in patients with extraprostatic extension or seminal vesicle invasion or lymph node involvement. It is interesting to note that in 50% of cases (5 of 10 cases) lymph node involvement indicated positive cav-1 expression. However, this difference did not reach statistical significance because of small sample size. These results, together with those of our previous study that demonstrated the frequency of cav-1 positivity was 8% in normal glandular epithelia, 29% in primary tumors with lymph node metastases, and 56% in the lymph node metastases per se,20 may indicate a association between cav-1 expression and the development of lymph node metastasis.
In the current study, univariate Cox proportional analysis was found to demonstrate that positive cav-1 expression predicted a shorter time to disease progression after radical prostatectomy in Japanese patients with clinically resectable prostate carcinoma. However, multivariate Cox proportional analysis demonstrated that positive cav-1 expression is not an independent prognostic marker. When we restricted the analysis to cases of pathologically confined prostate carcinoma (pT2N0), positive cav-1 was found to be the only parameter to predict time to disease recurrence.
Recent trends in prostate carcinoma diagnosis (e.g., lower serum PSA cutoff,11, 40 free-to-total PSA ratio,41–43 and complexed PSA44) have refined the indications for performing prostate biopsies. Increased utilization of prostate biopsies have, in turn, led to an increase in the number of men diagnosed with and treated for clinically localized prostate carcinoma. However, even when prostate carcinoma is pathologically confined to the gland, ≥ 14% of patients will develop biochemical or clinical recurrence within 5 years after surgery.45 The ability to predict disease recurrence in men with T2 disease has been studied by several groups. Blute et al. reported that positive surgical margins were a significant predictor of recurrence and were independent of tumor grade, serum PSA level, or DNA ploidy.45 Epstein et al. demonstrated that the Gleason score was the best predictor of disease progression and surgical margin was the only other variable that enhanced prediction, although it reportedly was less influential than grade.16 Conversely, Ohori et al. suggested that the prognosis of patients with positive surgical margins without extraprostatic extension was not different from that of patients with disease confined to the prostate.46 In 53 patients with large volume (≥ 6 cc), organ-confined tumors, only peripheral zone location and percent Gleason score 4/5 were found to be predictive of biochemical failure.19 It is interesting to note that preoperative serum PSA was not helpful in distinguishing biochemical failure rates in these large-volume tumors regardless of whether they were organ-confined.19 In our set of Japanese patients with pathologically confined prostate carcinoma, preoperative serum PSA, Gleason score, and positive surgical margins were not found to predict disease recurrence. Therefore, the use of molecular markers such as cav-1 may be a useful adjunct for predicting clinical outcome after surgery in Japanese patients with pathologically confined prostate carcinoma.
In the current study, we reported a positive correlation between cav-1 expression and surgical margin positivity; a higher incidence of cav-1 expression tended to be found in patients with poorly differentiated tumors (Gleason score > 7), extraprostatic extension, seminal vesicle invasion, or lymph node involvement. Univariate Cox proportional hazards analysis demonstrated that positive cav-1 expression was predictive of a shorter time to disease progression after radical prostatectomy. In this group of patients with pathologically determined, organ-confined disease (pT2N0), positive cav-1 expression was found to be the only significant parameter for predicting disease recurrence. Thus, cav-1 immunostaining in patients with prostate carcinoma may be a useful approach to predicting clinical outcome after surgery in Japanese patients with pT2N0 disease and may identify those patients who may benefit from novel adjuvant therapeutic strategies.
The authors thank Dr. Makoto Ohori, Dr. Takashi Saika, and Dr. Shin Ebara for their helpful discussions regarding the article.
- 5Prediction of cancer incidence in Japan. Tokyo: Shinohara Shuppan, 1999., , , .
- 34Evaluation of AxSYM automated immunoassay analyzer system for determining prostate-specific antigen, and comparison of reactivities for the antigen and diagnostic values of several commercial kits. J Med Pharm Sci. 1994; 32: 125–136., , .
- 37Regression models and life tables. J R Stat Soc. 1972; B34: 187–220..
- 41A complex between prostate-specific antigen and alpha 1-antichymotrypsin is the major form of prostate-specific antigen in serum of patients with prostatic cancer: assay of the complex improves clinical sensitivity for cancer. Cancer Res. 1991; 51: 222–226., , , , , .